![]() Investigate handling storage of request_token.oauth_token_secret instead of leaving that up to the dev.Code Cleanup! (Perhaps remove defapicall macro and use the with keyword instead? Have to evaluate.).Add models for label, checklist, member, notification, organization, session, token, action.member ( ) # Fetches the authenticated member record from Trello TODO: TokenAgent.retrieve("1234") from hypothetical TokenAgent = ExTrello. If you are able to invest the time there is almost no downside to building a side project.Oauth_token = "copied_oauth_token" oauth_verifier = "copied_oauth_verifier" oauth_token_secret = retrieve_oauth_token_secret_from_before ( ) # e.g. My point here is: There are advantages to starting a side project even if it never sees the light of the day. Of course, you can’t expect to get clients just based on Open Source contributions. I worked for 6 months in the development team using Elixir. We kept in contact and after some back and forth I was able to come on board as a freelancer. I learned he just started as the CTO of freshly founded startup. It turned out that he was living in the same city as I was. Halfway through abandoning the project, I received an email from the maintainer of the CouchDB library. I got a great insight into IBM Cloud Platform.I got to experience the limits of CouchDB views.I strengthened my knowledge in Docker, Event Sourcing.Was it a waste of time? I don’t think so. I designed myself into a corner (vaguely related to the fact that CouchDB views are not as versatile as I had hoped them to be)Ĭontinuing from here would have been very frustrating so I stopped the project.IBM’s clumsy interface had me frustrated.In the end, my side project failed for several reasons: ![]() I was lucky and after some back and forth my PR was accepted and I could continue to tinker with my project. I decided to go with option 1 and provide a pull request to the existing library. Will I amend the library, build my own data layer or give up on Cloudant completely? So here is where had to make another decision. I found out pretty quickly though that the library was lacking features I desperately needed. Nevertheless I decided to go with mkrogemann/couchdb_connector. Luckily CouchDB (and Cloudant) provide an HTTP based API, so it would have been pretty simple to refrain from using a library. There were a handful of libraries available for CouchDB. There are no dedicated libraries for available but luckily Cloudant is API compatible with Apache CouchDB. The toughest part, however, was getting Elixir to work together with Cloudant. I took an online course on ELM, read through the IBM Cloud Docs, built my own Docker-based CI pipeline to get my project deployed and of course, had to gather more knowledge in Event Sourcing and Phoenix. The Road Ahead #īecause I lacked experience in any of these technologies a lot of learning was involved. But as I am my own boss in this scenario I can do whatever I want. This stack is a pretty hard sell to management. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |